feat: add jina tools (#73)

tools/fetch_url_via_curl.sh

@@ -0,0 +1,16 @@
+#!/usr/bin/env bash
+set -e
+
+# @describe Extract the content from a given URL.
+# @meta require-tools pandoc
+# @option --url! The URL to scrape.
+
+main() {
+    # span and div tags are dropped from the HTML https://pandoc.org/MANUAL.html#raw-htmltex and sed removes any inline SVG images in image tags from the Markdown content.
+    curl -fsSL "$argc_url" | \
+        pandoc -f html-native_divs-native_spans -t gfm-raw_html | \
+        sed -E 's/!\[.*?\]\((data:image\/svg\+xml[^)]+)\)//g' \
+        >> "$LLM_OUTPUT"
+}
+
+eval "$(argc --argc-eval "$0" "$@")"

tools/fetch_url_via_jina.sh

@@ -0,0 +1,17 @@
+#!/usr/bin/env bash
+set -e
+
+# @describe Extract the content from a given URL.
+
+# @env JINA_API_KEY The api key
+# @option --url! The URL to scrape.
+
+main() {
+    curl_args=()
+    if [[ -n "$JINA_API_KEY" ]]; then
+        curl_args+=("-H" "Authorization: Bearer $JINA_API_KEY")
+    fi
+    curl -fsSL "${curl_args[@]}" "https://r.jina.ai/$argc_url" >> "$LLM_OUTPUT"
+}
+
+eval "$(argc --argc-eval "$0" "$@")"

tools/get_web_page.sh

@@ -1,16 +0,0 @@
-#!/usr/bin/env bash
-set -e
-
-# @describe Get webpage content from a given URL.
-# @meta require-tools pandoc
-# @option --url! The URL to scrape.
-
-main() {
-  # span and div tags are dropped from the HTML https://pandoc.org/MANUAL.html#raw-htmltex and sed removes any inline SVG images in image tags from the Markdown content.
-  curl -fsSL "$argc_url" | \
-    pandoc -f html-native_divs-native_spans -t gfm-raw_html | \
-    sed -E 's/!\[.*?\]\((data:image\/svg\+xml[^)]+)\)//g' \
-    >> "$LLM_OUTPUT"
-}
-
-eval "$(argc --argc-eval "$0" "$@")"

tools/search_jina.sh

@@ -0,0 +1,22 @@
+#!/usr/bin/env bash
+set -e
+
+# @describe Perform a web search using Jina API to get up-to-date information or additional context.
+# Use this when you need current information or feel a search could provide a better answer.
+
+# @env JINA_API_KEY The api key
+# @env JINA_MAX_RESULTS=5 The max results to return.
+# @option --query! The query to search for.
+
+main() {
+    curl_args=("-H" "Accept: application/json")
+    if [[ -n "$JINA_API_KEY" ]]; then
+        curl_args+=("-H" "Authorization: Bearer $JINA_API_KEY")
+    fi
+    encoded_query="$(jq -nr --arg q "$argc_query" '$q|@uri')"
+    curl -fsSL "${curl_args[@]}" "https://s.jina.ai/$encoded_query" | \
+        jq '[.data[:'"$JINA_MAX_RESULTS"'] | .[] | {title: .title, url: .url, description: .description}]' \
+        >> "$LLM_OUTPUT"
+}
+
+eval "$(argc --argc-eval "$0" "$@")"